🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Apple M3 Max 36GB run LUKE?
Yes — here's how fast and up to what context, with real measured numbers.
LUKE on the Apple M3 Max 36GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 457k | 211.2-254.7 t/s | ~7-8.5 (slow) |
| Q8 | yes | 464k | 422.4-509.3 t/s | ~14.1-17 (slow) |
| Q4 | yes | 467k | 754.2-909.5 t/s | ~25.1-30.3 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the Apple M3 Max 36GB run LUKE?▶
Yes. At Q4 it generates ~754.2-909.5 tok/s and fits up to 467k context; beyond that it spills to SSD swap and slows to ~25.1-30.3 tok/s.
How much context fits?▶
At Q4, up to about 467k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.